For a scanned batch that will be read by OCR and split at barcode separator sheets, scan at 300 dpi for normal office text (10 pt and larger), go to 400 dpi or more for small print, use grayscale rather than pure black and white when originals are faint or coloured, scan to PDF at actual size, and choose simplex or duplex deliberately so blank backs do not cause false splits. ScanRoute Local, a Windows app from Automata Labs, then splits the combined PDF at Code 128 or QR separator sheets, blank pages or every N pages, reads the date, client, type and number from one zone you draw, and shows every proposed file name before anything is written. Good scan settings are what make those readings right the first time.
Most copier defaults are tuned for small files, not for OCR. A few changes at the panel, made once, save a lot of typing in the exception inbox later.
Split a combined scan into named PDFs, with every proposed name shown before anything is written.
Get it from Microsoft
What resolution should you scan at for OCR?
OCR engines care about how many pixels tall each character is. Two vendor recommendations agree closely:
| Source | Recommendation |
|---|---|
| Nuance support | Best results when characters such as a lower-case “d” are 25 to 30 pixels tall; met by scanning a Letter page of 12 pt Times New Roman at 300 to 400 DPI |
| ABBYY support | 300 dpi for text of 10 pt and larger; 400 to 600 dpi for 9 pt and smaller; grayscale when there are shadows, texture or other noise, since black and white can lose character information on medium and low print quality |
In practice:
| Document | Suggested resolution |
|---|---|
| Letters, invoices, statements with normal body text | 300 dpi |
| Receipts, fine print, small table text | 400 dpi, or 600 dpi for thermal receipts with tiny type |
| Separator sheets only | 300 dpi is plenty for a well-printed barcode |
Going far above what the text needs mostly makes files larger and slower to process. Going below 300 dpi is the common cause of misread digits, such as 8 read as B or 0 as O.
Black and white, grayscale or colour?
| Mode | Use when | Watch for |
|---|---|---|
| Black and white (1-bit) | Clean, high-contrast originals | Faint text and light-coloured stamps can drop out entirely |
| Grayscale | Faint, coloured or low-contrast originals; ABBYY recommends it when shadows, texture or noise need removing | Larger files than black and white |
| Colour | Colour matters to the record, such as highlighted or signed documents | Largest files |
If the reference line you want read, such as an invoice number, is printed in a pale colour or sits on a tinted background, grayscale is the safer default. Some copiers apply strong compression in colour and grayscale modes to keep files small; if text edges look blocky when you zoom in, pick a higher-quality setting.
How should separator sheets be printed and scanned?
ScanRoute reads Code 128 and QR barcodes, one per sheet. The ScanRoute product page offers free printable separator sheets (prefixes SR-INV, SR-RCT and SR-CRD) with two instructions worth repeating:
- Print at actual size, not fit to page. Shrinking a barcode can make it harder to read.
- Scan a short test stack first. Five documents with separators show at once whether the splits land where you expect.
Set a separator prefix such as SR- in ScanRoute. The prefix match is case-sensitive, so a barcode printed on a document page, such as a purchase order, is not taken for a separator. Separator pages are left out of the output files. Keep separator sheets flat and clean; a creased or coffee-stained sheet that fails to read means two documents land in one file, which shows up in Preview and in the inbox as a missing separator.
Our guide to barcode separator sheets covers designing your own, and patch codes vs barcode separator sheets compares the two approaches.
Simplex or duplex?
Duplex scanning interacts with blank-page splitting, so decide which you are using:
| Splitting method | Simplex scanning | Duplex scanning |
|---|---|---|
| Barcode separators | Works as expected | Each separator’s blank back is also scanned; check in Preview how the stack splits, or print the separator on both sides |
| Blank pages | Works: insert a blank sheet between documents | Every single-sided page’s blank back looks like a split point; avoid unless your scanner removes blank backs |
| Every N pages | Works for fixed-length forms | N doubles: a 2-page form scanned duplex is 4 pages |
If your batches mix single- and double-sided documents, barcode separators with duplex scanning, tested on a short stack, are usually the most reliable. Our guide to splitting a PDF at blank pages covers that method in more detail.
What other copier settings matter?
- Output format: PDF. Scan to PDF, one file per batch, into a folder on the network or the PC. Our copier scan-to-folder guide covers the setup.
- Orientation. Feed pages upright. ScanRoute adds an invisible text layer only to upright scanned pages without text; rotated pages are skipped and noted in the manifest.
- Auto-crop and deskew, if the scanner offers them. Straight lines of text read better than skewed ones.
- No “OCR on the scanner” needed. ScanRoute reads text with the Windows OCR language installed on the PC. Scanner-side OCR is harmless, but not required.
How does ScanRoute use the scan once it lands?
| Step | What ScanRoute Local does |
|---|---|
| Collect | Drop combined PDFs on the window, open them, or turn on a watched folder; new PDFs wait until you plan them |
| Split | At Code 128 or QR separator sheets, at blank pages, or every N pages |
| Read | One zone you draw over the reference line; a pattern pulls fields such as type, date, client and number and checks each one |
| Name | A template builds each file name and folder from those fields |
| Preview | Every proposed name and folder listed before anything is written |
| Inbox | Reads that fail their checks (shown with the zone crop), missing separators, bad file names, duplicates, unreadable PDFs and no OCR language wait for a decision |
| Commit | Writes the files, with a CSV manifest listing each one and its SHA-256 hash; originals are never changed, moved or deleted, and no file is overwritten |
The exception inbox is where scan quality shows. If the same field keeps failing, look at the zone crop: blurry digits point to resolution, dropped-out text to black-and-white mode, and a zone in the wrong place to skew or a different form layout. Our guide to renaming scanned PDFs from an OCR zone explains how to place the zone.
Draw one zone, preview every name, and fix only the documents that need a decision.
Get it from Microsoft
A quick checklist for the copier panel
- Resolution 300 dpi (400 for small print).
- Grayscale for faint or coloured originals; black and white only for clean, dark text.
- Output PDF, one file per batch.
- Simplex, or duplex with barcode separators tested on a short stack.
- Separator sheets printed at actual size, prefix set in ScanRoute.
- Pages fed upright.
Save these as a one-touch preset on the copier if it allows, so everyone in the office scans the same way.
Frequently asked questions
Is 200 dpi enough for OCR?
For large, clean text it can work, but both Nuance and ABBYY recommend 300 dpi as the baseline for normal text. Misread digits in names and dates are the usual cost of scanning lower.
Should I scan in black and white to keep files small?
Only for clean, high-contrast originals. Faint text, coloured stamps and tinted backgrounds can drop out in black and white. ABBYY notes that black and white may lose character information and lower recognition quality on medium and low print quality documents.
Does ScanRoute need the scanner to do OCR?
No. ScanRoute reads text with the Windows OCR language installed on your PC. If none is installed, splitting still works and the fields wait for you to type them.
Can ScanRoute split at QR codes as well as Code 128?
Yes. It reads Code 128 and QR separator sheets, one barcode per sheet, matched by the prefix you set.
Does anything leave my PC?
No. ScanRoute Local makes no network connection, has no account and no telemetry.
Try ScanRoute Local
ScanRoute Local is a one-time purchase in the Microsoft Store with a free trial. Change the copier settings above, scan a short test stack with separator sheets, and see how many documents land in the inbox. More detail is on the ScanRoute product page, and the searchable scanned PDF page covers the text layer.
Test your scan settings on a short stack during the free trial.
Get it from Microsoft
Nuance, ABBYY and Windows are trademarks of their owners. ScanRoute Local is not affiliated with any scanner or OCR vendor.
Last updated 19 October 2026. Published by Automata Labs, a software company in Canada.